Automated Guided Vehicle Route Planning Problems Under Dynamic Environment

This paper mainly studies the Automated guided vehicle route planning problems under dynamic environment.These include the problem of multi-AGV route planning in the environment of moving obstacles,and the AGV lag problem that the actual route of the AGV does not match the planned route.For the first problem,this paper improves the traditional graph search algorithm——A * algorithm,and designs two algorithm which named Busy?A* and Time?A*.Both algorithms can solve the multi-AGV planning problem in a known dynamic environment.The former adds a busy heuristic function,allowing AGV to achieve collision-free by bypassing busy roads,and the operation speed is fast.The latter adds time dimension to the algorithm,changes the heuristic function and judges the deadlock in the dynamic environment which made the feasible solution better.For the second problem,this paper makes improvements to the Q-learning algorithm.First,the time dimension is added to make it possible to iterate over the dynamic environment step by step.Secondly,the prior search method was used to substitute the original exploration mode of the Q-learning algorithm.It can increase the search range of the environment for the algorithm,so that it can solve the lag problem.Finally,heuristic feedback value is designed to accelerate the convergence speed of the algorithm and solve the problem of dimensional disaster after adding the time dimension.In the simulation experiment module,a grid method is used to establish the environment model,and environments of different complexity are designed to simulate the designed algorithm.The advantages and disadvantages of the three algorithms and the applicable environment are explained.